[boost] [Property_tree review] Summary of enhancements proposed so far

From: Marcin Kalicinski (kalita_at_[hidden])
Date: 2006-04-22 08:48:43

I created a brief summary of enhancements/changes to property_tree which 
were proposed so far.

1. Make path a class in its own right and overload / (division) operator to 
concatenate path objects efficiently. This removes need for ugly default 
separator, and half of the get/put overloads. It also boosts performance (no 
more string operations on path), and IMHO will generally improve interface 
of the library. Most importantly, it largely does maintain compatibility 
with existing interface.

2. Split Traits into smaller policy classes. For example KeyPolicy and 
DataPolicy. Indeed, these two are mostly orthogonal, and it should be 
possible to replace one without touching the other.

3. Allow use of other data types as keys/paths (for example a vector of 
ints). Initial support is in the library, but key is still required to 
support string-like interface.

4. Add possibility to disable indexed lookup, if not required

5. Add possibility to specify names of "special" keys in parsers (<xmlattr>, 
<xmlcomment> etc.). Add whitespace-handling flags to XML reader.

6. Add "set" function. "set" will work in a similar fashion to put, except 
that it will always replace existing key if it exists. Extra boolean 
parameter to put functions will be gone. Alternative approach is to replace 
bool parameter to put with enum to improve readability.

7. Add "del" function. Del will erase existing keys from the tree, while 
presenting get/put/set style interface. This is clearly missing, although 
not immediately obvious.

8. Add support for array-style indexing in paths. This will probably work 
well with #1.

9. Add support for serialization of ptree objects

10. Add possibility to iterate over the whole tree in one go

11. Use multi_index internally for indexing

12. Make get/put functions freestanding

13. Add more docs on customizing key/data types.

14. Improve docs grammar/style in places. Break large index into smaller 
files.

Additionally there were proposals to use runtime polymorphism to distinguish 
between leaf and branch nodes. I think this idea strays so far from current 
implementation/interface of the library, that it is not feasible to do.

I hope this will serve as a basis for continued discussion.

Best regards,
Marcin